- Email: [email protected]
Gas-assisted compression moulding of recycled GMT: Effect of gas injection parameters
Journal of Materials Processing Technology 214 (2014) 515–523
Contents lists available at ScienceDirect
Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec
Gas-assisted compression moulding of recycled GMT: Effect of gas injection parameters V. Goodship ∗ , I. Brzeski, B.M. Wood, R. Cherrington, K. Makenji, N. Reynolds, G.J. Gibbons WMG, University of Warwick, Coventry CV4 7AL, UK
a r t i c l e
i n f o
Article history: Received 23 January 2013 Received in revised form 22 September 2013 Accepted 23 October 2013 Available online 31 October 2013 Keywords: Compression moulding Glass fibres Gas processing Interface Defects
a b s t r a c t Gas assisted compression moulding (or GasComp) is a novel process based on the injection of nitrogen gas during a conventional compression moulding cycle, a technique originally introduced in the injection moulding industry. The gas is injected into the molten material at a preset gas inlet point, hollowing out the part and thus assisting the compressive force of the press in generating material flow. This paper presents gas injection parameter studies on polypropylene based recycled glass mat-reinforced thermoplastic (GMT). The parameters investigated are gas ramp rate, gas injection delay time, and peak pressure. The size of cavity was found to be inversely proportional to gas injection delay time. Rheological instabilities at the polymer–gas interface were observed; a phenomenon previously noted during the development of gas assisted injection moulding. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Gas assisted injection moulding (GAIM) was first introduced in 1978 by Friedrich Ernst Dipling as a means to produce hollow components during the injection moulding process and has since been the subject of extensive research and commercialisation in order to obtain faster injection cycle times and better surface appearance on hollowed out components (Ernst, 1978). Avery (2001) provides useful coverage of developments up to 2000. The use of a cryogenically cooled gas as the hollowing media was developed by Magalhães (2002). It was found that using cryogenic gas led to reduced cycle times and smoother internal cavities when compared to the conventional GAIM process. The nitrogen gas has a lower density (1.165 kg/m3 ) than polypropylene (900 kg/m3 ) with the gas having a viscosity approximately 100 times lower than the polymer. There is, however, a gap in the literature regarding the use of gas injection to create similar benefits in the compression moulding process. Brzeski (2011) reported a novel process called gas assisted compression moulding (GasComp) where the use of gas in the compression moulding of various GMT materials was introduced. GasComp uses nitrogen gas to form an internal cavity in a glass mat-reinforced thermoplastic component during the compression moulding cycle. The compression moulding of glass fibre reinforced polymer, in combination with the internal cavity, enables high
∗ Corresponding author. Tel.: +44 02476 523684. E-mail address: [email protected] (V. Goodship). 0924-0136/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jmatprotec.2013.10.013
production levels of lightweight, load bearing components to be achieved. It was reported that in a comparison of the behaviour of four materials of various glass contents, glass alignment, fibre length and material flow rate the structure and orientation of the glass reinforcement were the major factor in material suitability for this process. Recycled GMT material produced the most consistent cavities with this process and exhibited more consistent gas penetration behaviour. Specifically it was reported shorter fibres (10 mm) produce more consistently shaped cavities than longer fibres (25 mm) or continuous fibres and that this process favours unidirectional glass configurations versus a crossed mat structure. It was also found the rheology of the material was less important than the fibre loading in terms of producing consistent cavities, a lower fibre loading producing more consistent results. This paper builds on previous work, by investigating the effect of the gas processing parameters on the recycled GMT material used in this study. This material has a glass loading of 11.8%, an average length of 10 mm and a melt flow rate of 1.35 g/10 min (2.16 kg at 200 ◦ C). The sequence of events for a typical GasComp moulding cycle is illustrated by Brzeski (2011), so will not be repeated here. The process is that a short shot of molten polymer composite material is moulded in a flash, fixed volume compression mould tool. In these previous trials, the shot size was set to fill approximately 80 percent of the cavity volume during the compression moulding stage. A gas pin integral to the mould tool then injects gas into the molten material, which creates a hollow in the polymer product, causing the displaced material to flow and fill the remaining volume of the
516
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
mould cavity. At a set point above full closure of the mould tool a gas injection delay time is triggered, after which a gas injection profile is followed. Once the gas profile and a variable vent time have been completed, any excess gas pressure is released. After the cooling time, the mould tool is opened and the part is removed. Combining compression moulding and gas injection presents a considerable technical challenge due to the number of processing parameters which can be varied in each of the two processes. With compression moulding Wakeman et al. (1999) carried out a Taguchi array experiment to establish the effects of different processing parameters on the mechanical properties of compression moulded GMT. It was found average GMT temperature, moulding pressure, time at pressure and compression velocity all had significant effects on mechanical properties. The time at pressure was concluded to have the most significant effect on the void content and therefore mechanical properties of the GMT laminate. Despite identifying these key parameters, the authors acknowledged that the literature contains conflicting recommendations on how they should be varied to improve the mechanical properties of compression moulded GMT. Whilst there are numerous recent papers using compression moulding and injection–compression moulding hybrids unfortunately none has further relevance to this research. To add to this complexity it is also necessary to consider the relevant gas processing parameters in GAIM. The parameters that can be controlled in GAIM in addition to conventional injection are short shot weight, gas delay time, peak gas pressure, pressure profile, packing pressure and vent time (gas exhaust time). Numerous studies have been carried out to determine the effects of various processing parameters on a moulded part. Shin (2002) uses gas bubble penetration (the distance the bubble travels in the polymer melt) and residual wall thickness (RWT) as measures of the effects on moulded components. These component parameters are the most common attributes by which the process settings changes are measured also being described by Becker et al. (1997) and Parvez et al. (2002), both of whose findings are illustrated further later in this paper. The reason for bubble penetration and RWT to be chosen is that they can be used to define the overall cavity volume. This is of particular interest in industry where controlling and predicting these attributes is important in predicting the physical properties of a component. These conventions will therefore be used here. The use of a pressurised gas to achieve complete filling of a mould is a common attribute of both GAIM of glass-reinforced polymers and the GasComp process. For example Becker et al. (1997) investigated the effect of glass weight fraction and fibre length on the gas channel appearance in a 12 mm diameter, 600 mm long spiral gas channel. Both polypropylene and polyamide matrix were used with different glass lengths and glass weight percentages. However, the processing parameters for this set of experiments were fixed for each matrix and not varied to take account of the different glass fibre lengths and weight fractions. The results allow a direct comparison of the matrices manufactured through both conventional and gas assisted injection moulding, however, the researchers did not optimise the process for the varying weight fractions and lengths of the glass fibres used. The investigation showed that all of the GAIM parts with glass lengths >4 mm, had some extent of fibre agglomeration or rough surface texture where glass rich areas occurred in the centre of the gas channel, within the first 75 mm length. The channel then generally alternated randomly between clear smooth sections and rougher, glass rich sections. The extent of these glass rich regions increased with increasing glass length and loading. Low delay times between the polymer injection and gas injection were needed to reduce the chance of the fibre agglomeration and aesthetic hesitation marks. A comparison of the effect of delay time in the GasComp process on glass filled materials is therefore of interest.
2. Materials and methods Given the range of variables (see Table 1) and the paucity of published research in this area, it was decided to study the material already reported to be most suitable for the GasComp process, namely recycled GMT (Re-GMT) and to explore the parameters affecting the gas injection phase; gas injection delay time, peak pressure and gas ramp rate. Unless explicitly stated all other parameters were kept constant throughout the testing process.
2.1. Material The recycled GMT (Re-GMT) used in this project was supplied by Voestalpine Polynorm Plastics. It was manufactured using slurry deposition, giving a dispersed glass fibre architecture and supplied in sheet form with a glass content of 20 percent by weight. (Note, this is a higher glass content than was previously reported for this process). The fibre length was approximately 10 mm. The effect of these various parameters on the moulding, dimensional stability and cavity volumes will be presented for a recycled GMT. To enable a basic rheological comparison for this material and those commonly encountered in injection moulding, melt mass-flow rate was measured to British Standard – ISO: 1133:1997 using “procedure A”. Flow rate was tested by chopping up the sheet into small pieces and feeding into a Ray Ran Melt Flow Indexer. This was measured at 1.35 g/10 min @ 2.16 kg and 200 ◦ C. When compared to melt flow ranges commonly encountered for simlarly loaded glass filled polypropylene injection moulding materials, this can be considered to be a much stiffer material than those previously investigated in injection moulding papers.
2.2. Loading method The dimension and weight of the initial blanks is given in Table 1, these were preheated in an infrared oven at 220 ◦ C for 6 min before being stacked on top of one another centrally in the tool cavity.
2.3. Machine and gas injection system The machine was a Dassett injection–compression machine, in which only the compression mode was utilised for this research. The press had a maximum clamp force 1000 kN. The gas injection system used was Airmold® which is manufactured by Wittman Battenfeld (Kottingbrunn, Austria). They also designed and supplied the gas injection pin which fitted into the tooling. The pin had a spring loaded shut off valve that kept it shut whilst it penetrated the material. The pin is forced open by the applied gas pressure, allowing the gas to flow into the material. This design prevents any ingress of material during the compression moulding cycle, which could block the pin and prevent any gas from reaching the material.
2.4. Tooling The compression moulding tool, illustrated in Fig. 1, was a flash plug design with a deep rib section and a horizontal clamping face, rather than a positive plug mould design with a vertical shear edge. A large hemisphere cavity was machined around the point where the gas injection pin would lie in the mould tool. This meant that a volume of material would be formed around the pin, providing a large mass of molten material through which the gas can flow. The large material mass would stay molten for a longer period, eliminating the chance of freeze off around the gas injection pin. The mould seal and gas pin assembly are shown in Figs. 2 and 3.
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
517
Table 1 Variables within the Gascomp process. Material
The temperature (◦ C, ±0.5 ◦ C) and time (s, ±0.1 s) at which the material was heated The dimensions of the initial material blanks. A number of blanks would be stacked on top of each other, in order to create the desired shot weight The shot weight dictates the volume of the material. Since the mould had a set volume, the shot weight would determine the maximum cavity potential (g, ±3 g)
Heating cycle Blank size
Shot weight
Press/mould tool
Stack Trigger position
Clamp force Cooling time Mould temperature Gas control
Delay time Gas profile Peak gas pressure Gas ramp Vent time
*
220 ◦ C for 6 min 200 mm × 105 mm
327 g
The way that the material was stacked together The point (mm, ±0.1 mm) in the press’s travel, which signalled the Air Mould Control unit – gas injection unit – to start its processing cycle. The trigger point was between 5 mm and 0 mm above the zero point (mould fully closed) The force (N, ±5 N) applied to the mould by the compression moulding press. The press had a maximum clamp force of 1000 kN The time (s, ±0.1 s) that the clamping force was applied for after the gas injection cycle had finished The directly measured temperature of the mould tool (◦ C, ±0.5 ◦ C)
Directly on top of each other 3 mm
The time delay (s, ±0.1 s) between the trigger position and the commencing of the gas profile A pre-selected profile of how the gas was injected. This was determined by a set gas pressure being reached by a set time The gas pressure (MPa, ±0.1 MPa) could be applied between 0.1 and 30 MPa, which was limited by the maximum system pressure. The time (s, ±0.1 s) at which the set gas pressure was required to have been reached Time delay (s, ±0.1 s) between the end of the gas injection cycle and the release of any trapped gas
*
Investigated in this paper.
Fig. 1. Technical drawing of the tool, (gas injection point is at the centre).
800 kN 300 s Ambient
*
*
*
20 s
518
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
Fig. 5. Gas pressure profiles for gas ramp experiments.
Fig. 2. Mould seal for gas pin.
Fig. 3. Gas Pin in assembly.
2.5. Experimental settings 2.5.1. Effect of gas injection delay time in Re-GMT The effect of 0.5 s, 1.5 s and 2.5 s gas injection delay times (as illustrated in Fig. 4) was investigated. Each of the mouldings had an identical gas pressure profile of reaching 80 bar over 1 s and this was held for 20 s. A large number of mouldings were produced during the trial, however, for clarity only three mouldings that displayed typical properties are presented in the results section.
2.5.2. Effect of gas ramp variation with Re-GMT The rate at which the gas pressure was increased was investigated. This was done by setting both a peak pressure and the time taken to reach it. The Airmould control unit then continually adjusts the flow of gas to achieve this goal. The gas injection had a delay time of 1 s and was held for 20 s in all experiments. Gas pressures below 40 bar and above 120 bar are not presented as no quantitative data was collected. It was found that gas pressures below 40 bar did not consistently produce a cavity and gas pressures above 120 bar consistently burst through the gas seal around the pin (the design of this seal is shown in Fig. 2). Process parameters are shown in Fig. 5. The gas ramp study used Re-GMT heated to 220 ◦ C for 6 min, in an unheated tool. 2.5.3. Moulded part characterisation A number of techniques were employed to characterise the moulded parts, which allowed them to be analysed in detail. The characterisation here is concentrated on the cavity formation. Measurement techniques used to calculate the cavity volumes were developed during the research; involving indirect and direct cavity measurements. Indirect cavity measurements used a combination of the materials density and the weight mass and volume of the moulded part, to calculate the cavity volume. Direct cavity measurement involved measuring the dimensions and volume directly. To retrieve the required information, the moulded part was sectioned and measured using accurate measuring equipment. Part dimensional stability and cavity measurements enabled an accurate comparison between parts and the effects of varying processing parameters. The measurements were performed using digital linear vernier callipers and a Hall effect probe to measure the wall thickness of the parts. These measurements were then imported into an excel spreadsheet for analysis. 3. Results and analysis 3.1. Re-GMT flow due to gas pressure
Fig. 4. Experimental gas profiles with differing delay times.
Each of the mouldings was sectioned and the internal cavity was then characterised using a Hall effect probe and vernier callipers. Each rib was measured at 10 mm intervals for the length of the cavity, width and height and this was used to calculate the total cavity volume. The volumes of each of the four ribs were then combined to give an overall part volume. These measurements for each of the delay times are shown in Fig. 6. Fig. 7 shows that the shapes of the cavities between the three delay times are similar, however the lengths are not. The moulding with a 0.5 s delay time, contains a fully cored out rib, whereas the other two mouldings do not. The moulding with a 2.5 s delay, contains more voids and breakthroughs. Breakthroughs are areas
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
Fig. 6. Cavity volumes with respect to gas delay time of each rib.
where the gas has penetrated areas of the component where a cavity was not intended. These findings show that the cavity wall thickness increases and cavity penetration decreases as the delay time increases up to 1.5 s. This is consistent with findings from injection moulding techniques produced by Parvez et al. (2002), stating that the delay time has a significant effect on the residual wall thickness (RWT). The total cavity volumes between 1.5 s and 2.5 s are very close at 24 cm3 and 23 cm3 , respectively. This would indicate that at some point between 0.5 s and 1.5 s significant frozen skin forms, which then insulates the rest of the polymer melt from the cold surface of the mould tool, meaning there is no significant change in wall thickness. This also indicates the potential importance of temperature gradients and residual polymer temperatures to this overall mechanism. From Fig. 6, the changes in the cavity volume of individual ribs are also of note. On further inspection of Rib 4, there appears to be more consistency in cavity volume here than in the other three ribs. It is not clear at this stage why this should be the case or if this is significant. However, all of the mouldings examined show a preference of gas flow with the clearest illustration in the moulding with a 0.5 gas delay time. This suggests the gas fully cores out one rib with another
Fig. 7. Picture of the cavities produced by a 0.5, 1.5 and 2.5 s delay time.
519
Fig. 8. Peak gas pressure (bar) vs cavity length (mm), with respect to gas ramp rate.
cavity length slightly less, whilst the two others are significantly shorter. This further indicates that the gas fully cored out the two larger cavities before flowing to the two shorter cavities until the part filled the mould tool. This could be attributed to a number of factors, inconsistencies within the gas injection point, tooling variations within the cavities, or temperature or glass content variations with the Re-GMT material itself. As gas delay increases from 0.5 to 1.5 s, the thickness of the frozen layer significantly increases, reducing the maximum potential cavity. The wall thickness between a moulding with a delay time of 1.5–2.5 s does not significantly change. The point at which the increase in delay time no longer significantly affects the cavity volume lies between 0.5 and 1.5 s. Further investigation of the delay time within this window would be valuable to further optimise the GasComp process for this material and tool combination. 3.2. Re-GMT flow due to peak pressure and ramp rate Fig. 8 shows the cavity lengths created by each peak pressure and ramp rate. The graph shows that as the ramp rate increases, the cavity length increases. The large variability in cavity length arises from the instability of the gas bubble in the main rib, the gas will always choose the path of least resistance. The reasons for the gas flow deviating away from the main rib could be due to the rib becoming full, localised freezing or a localised high concentration of glass fibres, which could create a dam, preventing the gas from advancing. If the sidewalls were still molten, the gas would then finger up the sidewalls to the top flange where the gas would either flow the material until the mould was full, flow the material until the gas burst through or flow the material until the material had sufficiently frozen as to contain the gas. Due to the unpredictable path, the extent of fingering could not be measured. Photographs taken from four samples at the shortest and longest ramp rates illustrate how the cavity sizes changes between the parameters, these are shown in Fig. 9. The photos show that the mouldings with the shorter ramp time have a shorter cavity length then the mouldings with a longer ramp rate. The cavity volumes of the mouldings shown in Fig. 9, as calculated theoretically, are displayed in Fig. 10. This shows a large variation in cavity volume within the same processing parameter. This is most likely due to slight variances in the shot sizes which results in variations in the fullness of the part and the burst through of the gas. The graph indicates that gas ramp and pressure do not significantly affect the volume of the cavity, giving a wide processing window for use. It also indicates that another parameter not investigated in this study had a larger affect on the cavity volume of the part, this is most likely a parameter of the compression moulding process itself in combination with the gas delay time
520
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
Fig. 9. Photo of moulding with various gas profiles, 60 bar over 1 s.
which at 1 s may still be affecting the cavity length. Further work would be required to clarify this further with a longer gas delay time of 1.5 s recommended. The cavities were characterised by a combination of methods: a vernier calliper, direct cavity measurement, a Hall effect probe and water displacement. Using a combination of methods a more accurate picture of all the cavities could be build up. The cavity dimensions and the standard deviation of those dimensions were taken along length of the rib. An increase of 2.4 mm in cavity height and 1.5 mm in cavity width was found from a gas profile of 60 bar over 1 s ramp rate up to a gas profile of 60 bar over 5 s ramp rate. The increase is not apparent once the gas pressure is increased to 80 bar. This shows that once the gas pressure reached 80 bar, the change in ramp rate did not affect the cavity height or width. Table 2 shows the change in the standard deviation of the cavity wall thickness with respect to gas ramp rate and gas pressure for 20 mouldings. The table shows a decrease in deviation as the gas ramp increases, showing that the cavity dimensions are more consistent and stable at a higher ramp rate. The gas ramp study appears to show that as the ramp rate increases, the cavity length increases. The total cavity volume itself does not change with gas pressure or gas ramp rate, suggesting that an alternative parameter outside the scope of this study has a major effect on the cavity volume. The higher ramp rate also appears to
Fig. 10. Graph showing the peak gas pressure (bar) vs CAD cavity volume (mm3 ), with respect to gas ramp rate.
produce a smoother, more consistent cavity wall thickness. This suggests rheology changes due to different temperature distribution fields at different gas ramping and therefore cooling rates are factors. A faster ramp rate to injection of gas would relate to higher polymer residual temperatures and therefore be able to penetrate further into the cavity length of the moulding before freeze off. Other factors could be inherent in the compression moulding process itself. 3.3. Further gas penetration investigations A further set of trials using the tool geometry shown in Figs. 11 and 12 were undertaken to gain further insight into the gas penetration mechanism. This tool produced just a single rib part of mass 218.01 g and volume of 222.45 cm3 . The component part produced is illustrated in Fig. 12 (right). Blanks were cut to the dimension of the tool and preheated in an oven for 6 min @ 220 ◦ C before placing them in the open cavity. With this set up the flow of Re-GMT mouldings through a cold rib processed with a 60 bar gas pressure over 1 s was further investigated. The mouldings were then prepared into 20 mm sections, a representative moulding is illustrated in Fig. 13 (top) and sections, one, five, nine, thirteen and seventeen, were polished and examined using an optical microscope. These sections were selected to correspond with the centre of the original blank, near the end of the original blank placement and at the end of the rib where there would be most compression flow of the material. Using multiple mouldings and sections the material flow due to the gas pressure and the progression of the gas cavity was investigated, this was completed by examining the gas cavity created in the Re-GMT moulding. Photos were taken of each of the sections in a single direction and processed to highlight the cavity shapes using a photoprocessing package. All 17 sections are displayed in Fig. 9 (bottom). The progression of the cavity can be seen to rapidly core out a significant portion of the rib, with some extent of fingering, before gradually reducing in size as it reaches the end of the rib. Section 2 shows how the gas fingered up into the top flange before expanding back towards the centre of the moulding. The same fingering effect can be seen at the opposite end of the moulding at section 14, where the gas expands in both directions along the top flange. Section 1 also shows that the gas was able to significantly core out the bottom flange. The gas will always follow the path of least resistance and this can change as the material flows and cools. The factors that decide
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
521
Table 2 Cavity dimensions and the standard deviation of the cavity height, width and overall area at each sampling point. Ramp time (s)
1 2 3 5 1 2 3 5
Peak pressure (bar)
60 60 60 60 80 80 80 80
Cavity dimensions (mm ±0.01 mm)
Cavity St. Dev.
Height
Width
Length
Height
Width
9.99 11.65 11.22 12.40 10.66 9.64 9.50 10.49
4.06 4.97 4.94 5.37 4.63 4.27 4.45 5.54
187.98 280.15 237.55 279.78 199.29 279.72 281.73 277.68
3.00 2.30 2.64 2.15 3.29 2.18 2.24 2.28
1.20 1.23 1.32 0.84 1.48 1.11 1.07 0.93
the path of least resistance are composite flow length, cross sectional area of cavity and composite temperature. The glass fibre may also present an obstacle, especially if glass fibres agglomerate under flow. Once the gas has reached the end of one flow path the gas pressure will then push the gas down an alternative flow path. The initial path could be decided by slight changes in the material characteristics such as blank positioning and localised increase of glass concentration, as the cross sectional area of the cavity and polymer temperature are uniform at the start of the cycle, prior to loading into the mould. The placement of the blanks in the mould would have the largest affect in the initial path of the gas by altering the flow length relative to the location of the gas injection pin. If the blank is offset to one side of the pin then this will have an increased resistance to flow due to the larger volume of material. The side with less material would have the lower resistance and so the gas would have a preference to travel in that direct first. This was evident in trials (not shown here) using blanks placed off centre.
Examination of mouldings show several gas mechanisms and a complex gas flow involving several stages as have been previously reported not just in fluid flow in injection moulding but also in co-injection moulding of rheologically mismatched materials. The work of Goodship and Kirwan (2001) provide a variety of illustrations of moulding defects resulting from rheological mismatch of skin and core, which is also evident here as the gas and GMT have significantly different rheological behaviours. The features that appear in Fig. 14, indicate that the gas took several paths at different time intervals in the injection. The gas injection profile used in this mould had a delay time of 1 s and a peak gas pressure of 120 bar over 1 s ramp rate. The polymer was assumed to have stopped flowing under compression before the injection of gas. The secondary flow front back into the frozen rib on the right hand side is not a common feature in other mouldings, as the gas would tend to finger up the side walls. This would indicate that the high gas pressure is the cause of this breakthrough. Another feature that is present in this moulding is entrainment. This is where high shear rate causes rheological instabilities on the surface of the polymer
Fig. 11. Single Rib tool.
522
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
Fig. 12. Rib tool and part produced (right).
Fig. 13. Moulding showing section positions and edge of the blank and subsequent compression moulding flow (top); Photos of sections 1–17 showing cavity shape progression in a rib moulding moulded with 60 bar peak pressure over 1 s (bottom).
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
523
and cavity length with an increasing trend up to a maximum gas pressure. This again can be attributed to rheological interactions. As the gas injection time increased, the cavity length also increased however this did not produce a higher cavity volume, the gas penetrated further into the moulding and the compression moulding flow front. Further parametric analysis is required to investigate variations in shot size, blank placement, temperature gradients and fibre alignment and to extend this study into gas injection delay times in order to further clarify this mechanism. Acknowledgments
Fig. 14. Close up of gas flow features; polymer breakthrough and entrainment.
that eventually close over, trapping the gas inside, again a similar phenomena is shown in Goodship and Kirwan (2001). Entrainment is more frequent at gas pressures above 80 bar. This is due to the rheological interactions of what is a skin–core mixture of materials, causing the gas to break through slower moving materials. There is further evidence of similar effect in the breakthrough region in Fig. 14. The cooling time has a direct influence on the crystalline structure of the material. The quicker the material is cooled, the less time the material has to form an ordered crystalline structure. As GAIM proves, a part with a reduced wall thickness and injected with cool nitrogen gas would cool quicker than solid material. 4. Conclusions This work has detailed a study of the gas assisted parameters on the gas assisted compression moulding of recycled GMT. This produced several major findings comparable to the processes also found in injection moulding including the formation of polymer–gas interface rheological instabilities. It was found a short gas injection delay time produces a larger cavity volume value than a longer delay time. This is attributed to cooling within the polymer affecting the rheology and therefore the ability of the gas to penetrate the polymer. There is also a relationship between gas pressure
The authors would like to thank the Engineering and Physical Sciences Research Council (UK) who funded the composite work and the European Regional Development Fund who supported the injection moulding expertise on this work. References Avery, J., 2001. Gas-Assist Injection Molding: Principles and Applications. Hanser Gardner Publication Inc, Cincinnati. Becker, O., Koelling, K., Altan, T., 1997. Gas-assisted injection molding of glass fiber reinforced thermoplastics. Journal of Injection Molding Technology 1 (3), 165–170. Brzeski, I., 2011. Gas-assisted compression moulding of glass reinforced polypropylene, part 1: scoping study. Progress in Rubber, Plastics and Recycling Technology 27 (2), 49–68. Ernst, F., 1978. In: European Patent Office E.P. Office (Ed.), Method for injection molding of hollow shaped bodies from thermoplastic resins. ROEHM GMBH. Goodship, V., Kirwan, K., 2001. Interfacial instabilities in multimaterial co-injection mouldings – Part 1 – background and initial experiments. Plastics, Rubber and Composites 30 (1), 11–15. Magalhães, R., 2002. Cryogenic gas-assisted injection moulding. School of Engineering, University Of Warwick, Coventry, UK (PhD thesis). Parvez, M.A., Ong, N.S., Lam, Y.C., Tor, S.B., 2002. Gas-assisted injection molding: the effects of process variables and gas channel geometry. Journal of Materials Processing Technology 121 (1), 27–35, http://dx.doi.org/10.1016/ S0924-0136(01)01184-0. Shin, J.-W., Isayev, A.I., 2002. Experimental study of gas penetration in gas-assisted injection molding. Journal of Injection Molding Technology 6 (4), 314–330. Wakeman, M.D., Cain, T.A., Rudd, C.D., Brooks, R., Long, A.C., 1999. Compression moulding of glass and polypropylene composites for optimised macro- and micro-mechanical properties II. Glass-mat-reinforced thermoplastics. Composites Science and Technology 59 (5), 709–726, http://dx.doi.org/ 10.1016/S0266-3538(98)00124-9.
Contents lists available at ScienceDirect
Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec
Gas-assisted compression moulding of recycled GMT: Effect of gas injection parameters V. Goodship ∗ , I. Brzeski, B.M. Wood, R. Cherrington, K. Makenji, N. Reynolds, G.J. Gibbons WMG, University of Warwick, Coventry CV4 7AL, UK
a r t i c l e
i n f o
Article history: Received 23 January 2013 Received in revised form 22 September 2013 Accepted 23 October 2013 Available online 31 October 2013 Keywords: Compression moulding Glass fibres Gas processing Interface Defects
a b s t r a c t Gas assisted compression moulding (or GasComp) is a novel process based on the injection of nitrogen gas during a conventional compression moulding cycle, a technique originally introduced in the injection moulding industry. The gas is injected into the molten material at a preset gas inlet point, hollowing out the part and thus assisting the compressive force of the press in generating material flow. This paper presents gas injection parameter studies on polypropylene based recycled glass mat-reinforced thermoplastic (GMT). The parameters investigated are gas ramp rate, gas injection delay time, and peak pressure. The size of cavity was found to be inversely proportional to gas injection delay time. Rheological instabilities at the polymer–gas interface were observed; a phenomenon previously noted during the development of gas assisted injection moulding. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Gas assisted injection moulding (GAIM) was first introduced in 1978 by Friedrich Ernst Dipling as a means to produce hollow components during the injection moulding process and has since been the subject of extensive research and commercialisation in order to obtain faster injection cycle times and better surface appearance on hollowed out components (Ernst, 1978). Avery (2001) provides useful coverage of developments up to 2000. The use of a cryogenically cooled gas as the hollowing media was developed by Magalhães (2002). It was found that using cryogenic gas led to reduced cycle times and smoother internal cavities when compared to the conventional GAIM process. The nitrogen gas has a lower density (1.165 kg/m3 ) than polypropylene (900 kg/m3 ) with the gas having a viscosity approximately 100 times lower than the polymer. There is, however, a gap in the literature regarding the use of gas injection to create similar benefits in the compression moulding process. Brzeski (2011) reported a novel process called gas assisted compression moulding (GasComp) where the use of gas in the compression moulding of various GMT materials was introduced. GasComp uses nitrogen gas to form an internal cavity in a glass mat-reinforced thermoplastic component during the compression moulding cycle. The compression moulding of glass fibre reinforced polymer, in combination with the internal cavity, enables high
∗ Corresponding author. Tel.: +44 02476 523684. E-mail address: [email protected] (V. Goodship). 0924-0136/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jmatprotec.2013.10.013
production levels of lightweight, load bearing components to be achieved. It was reported that in a comparison of the behaviour of four materials of various glass contents, glass alignment, fibre length and material flow rate the structure and orientation of the glass reinforcement were the major factor in material suitability for this process. Recycled GMT material produced the most consistent cavities with this process and exhibited more consistent gas penetration behaviour. Specifically it was reported shorter fibres (10 mm) produce more consistently shaped cavities than longer fibres (25 mm) or continuous fibres and that this process favours unidirectional glass configurations versus a crossed mat structure. It was also found the rheology of the material was less important than the fibre loading in terms of producing consistent cavities, a lower fibre loading producing more consistent results. This paper builds on previous work, by investigating the effect of the gas processing parameters on the recycled GMT material used in this study. This material has a glass loading of 11.8%, an average length of 10 mm and a melt flow rate of 1.35 g/10 min (2.16 kg at 200 ◦ C). The sequence of events for a typical GasComp moulding cycle is illustrated by Brzeski (2011), so will not be repeated here. The process is that a short shot of molten polymer composite material is moulded in a flash, fixed volume compression mould tool. In these previous trials, the shot size was set to fill approximately 80 percent of the cavity volume during the compression moulding stage. A gas pin integral to the mould tool then injects gas into the molten material, which creates a hollow in the polymer product, causing the displaced material to flow and fill the remaining volume of the
516
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
mould cavity. At a set point above full closure of the mould tool a gas injection delay time is triggered, after which a gas injection profile is followed. Once the gas profile and a variable vent time have been completed, any excess gas pressure is released. After the cooling time, the mould tool is opened and the part is removed. Combining compression moulding and gas injection presents a considerable technical challenge due to the number of processing parameters which can be varied in each of the two processes. With compression moulding Wakeman et al. (1999) carried out a Taguchi array experiment to establish the effects of different processing parameters on the mechanical properties of compression moulded GMT. It was found average GMT temperature, moulding pressure, time at pressure and compression velocity all had significant effects on mechanical properties. The time at pressure was concluded to have the most significant effect on the void content and therefore mechanical properties of the GMT laminate. Despite identifying these key parameters, the authors acknowledged that the literature contains conflicting recommendations on how they should be varied to improve the mechanical properties of compression moulded GMT. Whilst there are numerous recent papers using compression moulding and injection–compression moulding hybrids unfortunately none has further relevance to this research. To add to this complexity it is also necessary to consider the relevant gas processing parameters in GAIM. The parameters that can be controlled in GAIM in addition to conventional injection are short shot weight, gas delay time, peak gas pressure, pressure profile, packing pressure and vent time (gas exhaust time). Numerous studies have been carried out to determine the effects of various processing parameters on a moulded part. Shin (2002) uses gas bubble penetration (the distance the bubble travels in the polymer melt) and residual wall thickness (RWT) as measures of the effects on moulded components. These component parameters are the most common attributes by which the process settings changes are measured also being described by Becker et al. (1997) and Parvez et al. (2002), both of whose findings are illustrated further later in this paper. The reason for bubble penetration and RWT to be chosen is that they can be used to define the overall cavity volume. This is of particular interest in industry where controlling and predicting these attributes is important in predicting the physical properties of a component. These conventions will therefore be used here. The use of a pressurised gas to achieve complete filling of a mould is a common attribute of both GAIM of glass-reinforced polymers and the GasComp process. For example Becker et al. (1997) investigated the effect of glass weight fraction and fibre length on the gas channel appearance in a 12 mm diameter, 600 mm long spiral gas channel. Both polypropylene and polyamide matrix were used with different glass lengths and glass weight percentages. However, the processing parameters for this set of experiments were fixed for each matrix and not varied to take account of the different glass fibre lengths and weight fractions. The results allow a direct comparison of the matrices manufactured through both conventional and gas assisted injection moulding, however, the researchers did not optimise the process for the varying weight fractions and lengths of the glass fibres used. The investigation showed that all of the GAIM parts with glass lengths >4 mm, had some extent of fibre agglomeration or rough surface texture where glass rich areas occurred in the centre of the gas channel, within the first 75 mm length. The channel then generally alternated randomly between clear smooth sections and rougher, glass rich sections. The extent of these glass rich regions increased with increasing glass length and loading. Low delay times between the polymer injection and gas injection were needed to reduce the chance of the fibre agglomeration and aesthetic hesitation marks. A comparison of the effect of delay time in the GasComp process on glass filled materials is therefore of interest.
2. Materials and methods Given the range of variables (see Table 1) and the paucity of published research in this area, it was decided to study the material already reported to be most suitable for the GasComp process, namely recycled GMT (Re-GMT) and to explore the parameters affecting the gas injection phase; gas injection delay time, peak pressure and gas ramp rate. Unless explicitly stated all other parameters were kept constant throughout the testing process.
2.1. Material The recycled GMT (Re-GMT) used in this project was supplied by Voestalpine Polynorm Plastics. It was manufactured using slurry deposition, giving a dispersed glass fibre architecture and supplied in sheet form with a glass content of 20 percent by weight. (Note, this is a higher glass content than was previously reported for this process). The fibre length was approximately 10 mm. The effect of these various parameters on the moulding, dimensional stability and cavity volumes will be presented for a recycled GMT. To enable a basic rheological comparison for this material and those commonly encountered in injection moulding, melt mass-flow rate was measured to British Standard – ISO: 1133:1997 using “procedure A”. Flow rate was tested by chopping up the sheet into small pieces and feeding into a Ray Ran Melt Flow Indexer. This was measured at 1.35 g/10 min @ 2.16 kg and 200 ◦ C. When compared to melt flow ranges commonly encountered for simlarly loaded glass filled polypropylene injection moulding materials, this can be considered to be a much stiffer material than those previously investigated in injection moulding papers.
2.2. Loading method The dimension and weight of the initial blanks is given in Table 1, these were preheated in an infrared oven at 220 ◦ C for 6 min before being stacked on top of one another centrally in the tool cavity.
2.3. Machine and gas injection system The machine was a Dassett injection–compression machine, in which only the compression mode was utilised for this research. The press had a maximum clamp force 1000 kN. The gas injection system used was Airmold® which is manufactured by Wittman Battenfeld (Kottingbrunn, Austria). They also designed and supplied the gas injection pin which fitted into the tooling. The pin had a spring loaded shut off valve that kept it shut whilst it penetrated the material. The pin is forced open by the applied gas pressure, allowing the gas to flow into the material. This design prevents any ingress of material during the compression moulding cycle, which could block the pin and prevent any gas from reaching the material.
2.4. Tooling The compression moulding tool, illustrated in Fig. 1, was a flash plug design with a deep rib section and a horizontal clamping face, rather than a positive plug mould design with a vertical shear edge. A large hemisphere cavity was machined around the point where the gas injection pin would lie in the mould tool. This meant that a volume of material would be formed around the pin, providing a large mass of molten material through which the gas can flow. The large material mass would stay molten for a longer period, eliminating the chance of freeze off around the gas injection pin. The mould seal and gas pin assembly are shown in Figs. 2 and 3.
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
517
Table 1 Variables within the Gascomp process. Material
The temperature (◦ C, ±0.5 ◦ C) and time (s, ±0.1 s) at which the material was heated The dimensions of the initial material blanks. A number of blanks would be stacked on top of each other, in order to create the desired shot weight The shot weight dictates the volume of the material. Since the mould had a set volume, the shot weight would determine the maximum cavity potential (g, ±3 g)
Heating cycle Blank size
Shot weight
Press/mould tool
Stack Trigger position
Clamp force Cooling time Mould temperature Gas control
Delay time Gas profile Peak gas pressure Gas ramp Vent time
*
220 ◦ C for 6 min 200 mm × 105 mm
327 g
The way that the material was stacked together The point (mm, ±0.1 mm) in the press’s travel, which signalled the Air Mould Control unit – gas injection unit – to start its processing cycle. The trigger point was between 5 mm and 0 mm above the zero point (mould fully closed) The force (N, ±5 N) applied to the mould by the compression moulding press. The press had a maximum clamp force of 1000 kN The time (s, ±0.1 s) that the clamping force was applied for after the gas injection cycle had finished The directly measured temperature of the mould tool (◦ C, ±0.5 ◦ C)
Directly on top of each other 3 mm
The time delay (s, ±0.1 s) between the trigger position and the commencing of the gas profile A pre-selected profile of how the gas was injected. This was determined by a set gas pressure being reached by a set time The gas pressure (MPa, ±0.1 MPa) could be applied between 0.1 and 30 MPa, which was limited by the maximum system pressure. The time (s, ±0.1 s) at which the set gas pressure was required to have been reached Time delay (s, ±0.1 s) between the end of the gas injection cycle and the release of any trapped gas
*
Investigated in this paper.
Fig. 1. Technical drawing of the tool, (gas injection point is at the centre).
800 kN 300 s Ambient
*
*
*
20 s
518
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
Fig. 5. Gas pressure profiles for gas ramp experiments.
Fig. 2. Mould seal for gas pin.
Fig. 3. Gas Pin in assembly.
2.5. Experimental settings 2.5.1. Effect of gas injection delay time in Re-GMT The effect of 0.5 s, 1.5 s and 2.5 s gas injection delay times (as illustrated in Fig. 4) was investigated. Each of the mouldings had an identical gas pressure profile of reaching 80 bar over 1 s and this was held for 20 s. A large number of mouldings were produced during the trial, however, for clarity only three mouldings that displayed typical properties are presented in the results section.
2.5.2. Effect of gas ramp variation with Re-GMT The rate at which the gas pressure was increased was investigated. This was done by setting both a peak pressure and the time taken to reach it. The Airmould control unit then continually adjusts the flow of gas to achieve this goal. The gas injection had a delay time of 1 s and was held for 20 s in all experiments. Gas pressures below 40 bar and above 120 bar are not presented as no quantitative data was collected. It was found that gas pressures below 40 bar did not consistently produce a cavity and gas pressures above 120 bar consistently burst through the gas seal around the pin (the design of this seal is shown in Fig. 2). Process parameters are shown in Fig. 5. The gas ramp study used Re-GMT heated to 220 ◦ C for 6 min, in an unheated tool. 2.5.3. Moulded part characterisation A number of techniques were employed to characterise the moulded parts, which allowed them to be analysed in detail. The characterisation here is concentrated on the cavity formation. Measurement techniques used to calculate the cavity volumes were developed during the research; involving indirect and direct cavity measurements. Indirect cavity measurements used a combination of the materials density and the weight mass and volume of the moulded part, to calculate the cavity volume. Direct cavity measurement involved measuring the dimensions and volume directly. To retrieve the required information, the moulded part was sectioned and measured using accurate measuring equipment. Part dimensional stability and cavity measurements enabled an accurate comparison between parts and the effects of varying processing parameters. The measurements were performed using digital linear vernier callipers and a Hall effect probe to measure the wall thickness of the parts. These measurements were then imported into an excel spreadsheet for analysis. 3. Results and analysis 3.1. Re-GMT flow due to gas pressure
Fig. 4. Experimental gas profiles with differing delay times.
Each of the mouldings was sectioned and the internal cavity was then characterised using a Hall effect probe and vernier callipers. Each rib was measured at 10 mm intervals for the length of the cavity, width and height and this was used to calculate the total cavity volume. The volumes of each of the four ribs were then combined to give an overall part volume. These measurements for each of the delay times are shown in Fig. 6. Fig. 7 shows that the shapes of the cavities between the three delay times are similar, however the lengths are not. The moulding with a 0.5 s delay time, contains a fully cored out rib, whereas the other two mouldings do not. The moulding with a 2.5 s delay, contains more voids and breakthroughs. Breakthroughs are areas
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
Fig. 6. Cavity volumes with respect to gas delay time of each rib.
where the gas has penetrated areas of the component where a cavity was not intended. These findings show that the cavity wall thickness increases and cavity penetration decreases as the delay time increases up to 1.5 s. This is consistent with findings from injection moulding techniques produced by Parvez et al. (2002), stating that the delay time has a significant effect on the residual wall thickness (RWT). The total cavity volumes between 1.5 s and 2.5 s are very close at 24 cm3 and 23 cm3 , respectively. This would indicate that at some point between 0.5 s and 1.5 s significant frozen skin forms, which then insulates the rest of the polymer melt from the cold surface of the mould tool, meaning there is no significant change in wall thickness. This also indicates the potential importance of temperature gradients and residual polymer temperatures to this overall mechanism. From Fig. 6, the changes in the cavity volume of individual ribs are also of note. On further inspection of Rib 4, there appears to be more consistency in cavity volume here than in the other three ribs. It is not clear at this stage why this should be the case or if this is significant. However, all of the mouldings examined show a preference of gas flow with the clearest illustration in the moulding with a 0.5 gas delay time. This suggests the gas fully cores out one rib with another
Fig. 7. Picture of the cavities produced by a 0.5, 1.5 and 2.5 s delay time.
519
Fig. 8. Peak gas pressure (bar) vs cavity length (mm), with respect to gas ramp rate.
cavity length slightly less, whilst the two others are significantly shorter. This further indicates that the gas fully cored out the two larger cavities before flowing to the two shorter cavities until the part filled the mould tool. This could be attributed to a number of factors, inconsistencies within the gas injection point, tooling variations within the cavities, or temperature or glass content variations with the Re-GMT material itself. As gas delay increases from 0.5 to 1.5 s, the thickness of the frozen layer significantly increases, reducing the maximum potential cavity. The wall thickness between a moulding with a delay time of 1.5–2.5 s does not significantly change. The point at which the increase in delay time no longer significantly affects the cavity volume lies between 0.5 and 1.5 s. Further investigation of the delay time within this window would be valuable to further optimise the GasComp process for this material and tool combination. 3.2. Re-GMT flow due to peak pressure and ramp rate Fig. 8 shows the cavity lengths created by each peak pressure and ramp rate. The graph shows that as the ramp rate increases, the cavity length increases. The large variability in cavity length arises from the instability of the gas bubble in the main rib, the gas will always choose the path of least resistance. The reasons for the gas flow deviating away from the main rib could be due to the rib becoming full, localised freezing or a localised high concentration of glass fibres, which could create a dam, preventing the gas from advancing. If the sidewalls were still molten, the gas would then finger up the sidewalls to the top flange where the gas would either flow the material until the mould was full, flow the material until the gas burst through or flow the material until the material had sufficiently frozen as to contain the gas. Due to the unpredictable path, the extent of fingering could not be measured. Photographs taken from four samples at the shortest and longest ramp rates illustrate how the cavity sizes changes between the parameters, these are shown in Fig. 9. The photos show that the mouldings with the shorter ramp time have a shorter cavity length then the mouldings with a longer ramp rate. The cavity volumes of the mouldings shown in Fig. 9, as calculated theoretically, are displayed in Fig. 10. This shows a large variation in cavity volume within the same processing parameter. This is most likely due to slight variances in the shot sizes which results in variations in the fullness of the part and the burst through of the gas. The graph indicates that gas ramp and pressure do not significantly affect the volume of the cavity, giving a wide processing window for use. It also indicates that another parameter not investigated in this study had a larger affect on the cavity volume of the part, this is most likely a parameter of the compression moulding process itself in combination with the gas delay time
520
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
Fig. 9. Photo of moulding with various gas profiles, 60 bar over 1 s.
which at 1 s may still be affecting the cavity length. Further work would be required to clarify this further with a longer gas delay time of 1.5 s recommended. The cavities were characterised by a combination of methods: a vernier calliper, direct cavity measurement, a Hall effect probe and water displacement. Using a combination of methods a more accurate picture of all the cavities could be build up. The cavity dimensions and the standard deviation of those dimensions were taken along length of the rib. An increase of 2.4 mm in cavity height and 1.5 mm in cavity width was found from a gas profile of 60 bar over 1 s ramp rate up to a gas profile of 60 bar over 5 s ramp rate. The increase is not apparent once the gas pressure is increased to 80 bar. This shows that once the gas pressure reached 80 bar, the change in ramp rate did not affect the cavity height or width. Table 2 shows the change in the standard deviation of the cavity wall thickness with respect to gas ramp rate and gas pressure for 20 mouldings. The table shows a decrease in deviation as the gas ramp increases, showing that the cavity dimensions are more consistent and stable at a higher ramp rate. The gas ramp study appears to show that as the ramp rate increases, the cavity length increases. The total cavity volume itself does not change with gas pressure or gas ramp rate, suggesting that an alternative parameter outside the scope of this study has a major effect on the cavity volume. The higher ramp rate also appears to
Fig. 10. Graph showing the peak gas pressure (bar) vs CAD cavity volume (mm3 ), with respect to gas ramp rate.
produce a smoother, more consistent cavity wall thickness. This suggests rheology changes due to different temperature distribution fields at different gas ramping and therefore cooling rates are factors. A faster ramp rate to injection of gas would relate to higher polymer residual temperatures and therefore be able to penetrate further into the cavity length of the moulding before freeze off. Other factors could be inherent in the compression moulding process itself. 3.3. Further gas penetration investigations A further set of trials using the tool geometry shown in Figs. 11 and 12 were undertaken to gain further insight into the gas penetration mechanism. This tool produced just a single rib part of mass 218.01 g and volume of 222.45 cm3 . The component part produced is illustrated in Fig. 12 (right). Blanks were cut to the dimension of the tool and preheated in an oven for 6 min @ 220 ◦ C before placing them in the open cavity. With this set up the flow of Re-GMT mouldings through a cold rib processed with a 60 bar gas pressure over 1 s was further investigated. The mouldings were then prepared into 20 mm sections, a representative moulding is illustrated in Fig. 13 (top) and sections, one, five, nine, thirteen and seventeen, were polished and examined using an optical microscope. These sections were selected to correspond with the centre of the original blank, near the end of the original blank placement and at the end of the rib where there would be most compression flow of the material. Using multiple mouldings and sections the material flow due to the gas pressure and the progression of the gas cavity was investigated, this was completed by examining the gas cavity created in the Re-GMT moulding. Photos were taken of each of the sections in a single direction and processed to highlight the cavity shapes using a photoprocessing package. All 17 sections are displayed in Fig. 9 (bottom). The progression of the cavity can be seen to rapidly core out a significant portion of the rib, with some extent of fingering, before gradually reducing in size as it reaches the end of the rib. Section 2 shows how the gas fingered up into the top flange before expanding back towards the centre of the moulding. The same fingering effect can be seen at the opposite end of the moulding at section 14, where the gas expands in both directions along the top flange. Section 1 also shows that the gas was able to significantly core out the bottom flange. The gas will always follow the path of least resistance and this can change as the material flows and cools. The factors that decide
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
521
Table 2 Cavity dimensions and the standard deviation of the cavity height, width and overall area at each sampling point. Ramp time (s)
1 2 3 5 1 2 3 5
Peak pressure (bar)
60 60 60 60 80 80 80 80
Cavity dimensions (mm ±0.01 mm)
Cavity St. Dev.
Height
Width
Length
Height
Width
9.99 11.65 11.22 12.40 10.66 9.64 9.50 10.49
4.06 4.97 4.94 5.37 4.63 4.27 4.45 5.54
187.98 280.15 237.55 279.78 199.29 279.72 281.73 277.68
3.00 2.30 2.64 2.15 3.29 2.18 2.24 2.28
1.20 1.23 1.32 0.84 1.48 1.11 1.07 0.93
the path of least resistance are composite flow length, cross sectional area of cavity and composite temperature. The glass fibre may also present an obstacle, especially if glass fibres agglomerate under flow. Once the gas has reached the end of one flow path the gas pressure will then push the gas down an alternative flow path. The initial path could be decided by slight changes in the material characteristics such as blank positioning and localised increase of glass concentration, as the cross sectional area of the cavity and polymer temperature are uniform at the start of the cycle, prior to loading into the mould. The placement of the blanks in the mould would have the largest affect in the initial path of the gas by altering the flow length relative to the location of the gas injection pin. If the blank is offset to one side of the pin then this will have an increased resistance to flow due to the larger volume of material. The side with less material would have the lower resistance and so the gas would have a preference to travel in that direct first. This was evident in trials (not shown here) using blanks placed off centre.
Examination of mouldings show several gas mechanisms and a complex gas flow involving several stages as have been previously reported not just in fluid flow in injection moulding but also in co-injection moulding of rheologically mismatched materials. The work of Goodship and Kirwan (2001) provide a variety of illustrations of moulding defects resulting from rheological mismatch of skin and core, which is also evident here as the gas and GMT have significantly different rheological behaviours. The features that appear in Fig. 14, indicate that the gas took several paths at different time intervals in the injection. The gas injection profile used in this mould had a delay time of 1 s and a peak gas pressure of 120 bar over 1 s ramp rate. The polymer was assumed to have stopped flowing under compression before the injection of gas. The secondary flow front back into the frozen rib on the right hand side is not a common feature in other mouldings, as the gas would tend to finger up the side walls. This would indicate that the high gas pressure is the cause of this breakthrough. Another feature that is present in this moulding is entrainment. This is where high shear rate causes rheological instabilities on the surface of the polymer
Fig. 11. Single Rib tool.
522
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
Fig. 12. Rib tool and part produced (right).
Fig. 13. Moulding showing section positions and edge of the blank and subsequent compression moulding flow (top); Photos of sections 1–17 showing cavity shape progression in a rib moulding moulded with 60 bar peak pressure over 1 s (bottom).
V. Goodship et al. / Journal of Materials Processing Technology 214 (2014) 515–523
523
and cavity length with an increasing trend up to a maximum gas pressure. This again can be attributed to rheological interactions. As the gas injection time increased, the cavity length also increased however this did not produce a higher cavity volume, the gas penetrated further into the moulding and the compression moulding flow front. Further parametric analysis is required to investigate variations in shot size, blank placement, temperature gradients and fibre alignment and to extend this study into gas injection delay times in order to further clarify this mechanism. Acknowledgments
Fig. 14. Close up of gas flow features; polymer breakthrough and entrainment.
that eventually close over, trapping the gas inside, again a similar phenomena is shown in Goodship and Kirwan (2001). Entrainment is more frequent at gas pressures above 80 bar. This is due to the rheological interactions of what is a skin–core mixture of materials, causing the gas to break through slower moving materials. There is further evidence of similar effect in the breakthrough region in Fig. 14. The cooling time has a direct influence on the crystalline structure of the material. The quicker the material is cooled, the less time the material has to form an ordered crystalline structure. As GAIM proves, a part with a reduced wall thickness and injected with cool nitrogen gas would cool quicker than solid material. 4. Conclusions This work has detailed a study of the gas assisted parameters on the gas assisted compression moulding of recycled GMT. This produced several major findings comparable to the processes also found in injection moulding including the formation of polymer–gas interface rheological instabilities. It was found a short gas injection delay time produces a larger cavity volume value than a longer delay time. This is attributed to cooling within the polymer affecting the rheology and therefore the ability of the gas to penetrate the polymer. There is also a relationship between gas pressure
The authors would like to thank the Engineering and Physical Sciences Research Council (UK) who funded the composite work and the European Regional Development Fund who supported the injection moulding expertise on this work. References Avery, J., 2001. Gas-Assist Injection Molding: Principles and Applications. Hanser Gardner Publication Inc, Cincinnati. Becker, O., Koelling, K., Altan, T., 1997. Gas-assisted injection molding of glass fiber reinforced thermoplastics. Journal of Injection Molding Technology 1 (3), 165–170. Brzeski, I., 2011. Gas-assisted compression moulding of glass reinforced polypropylene, part 1: scoping study. Progress in Rubber, Plastics and Recycling Technology 27 (2), 49–68. Ernst, F., 1978. In: European Patent Office E.P. Office (Ed.), Method for injection molding of hollow shaped bodies from thermoplastic resins. ROEHM GMBH. Goodship, V., Kirwan, K., 2001. Interfacial instabilities in multimaterial co-injection mouldings – Part 1 – background and initial experiments. Plastics, Rubber and Composites 30 (1), 11–15. Magalhães, R., 2002. Cryogenic gas-assisted injection moulding. School of Engineering, University Of Warwick, Coventry, UK (PhD thesis). Parvez, M.A., Ong, N.S., Lam, Y.C., Tor, S.B., 2002. Gas-assisted injection molding: the effects of process variables and gas channel geometry. Journal of Materials Processing Technology 121 (1), 27–35, http://dx.doi.org/10.1016/ S0924-0136(01)01184-0. Shin, J.-W., Isayev, A.I., 2002. Experimental study of gas penetration in gas-assisted injection molding. Journal of Injection Molding Technology 6 (4), 314–330. Wakeman, M.D., Cain, T.A., Rudd, C.D., Brooks, R., Long, A.C., 1999. Compression moulding of glass and polypropylene composites for optimised macro- and micro-mechanical properties II. Glass-mat-reinforced thermoplastics. Composites Science and Technology 59 (5), 709–726, http://dx.doi.org/ 10.1016/S0266-3538(98)00124-9.